A Workflow to Accurately Delineate Implanted hardware in Proton Planning
F Su*, A Paxton, S St. James, X Li, R Price, B Salter, V Sarkar, University of Utah - Huntsman Cancer Institute, Salt Lake City, UT
Presentations
PO-GePV-T-176 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: To improve the accuracy of modeling implanted hardware in proton planning by use of a stereolithographic (STL) approach.
Methods: We used two approaches to create STL models of implanted hardware with the computer-aided design (CAD) software, Autodesk Fusion 360. One way to accurately model the hardware is to sketch it from a two-dimensional (2D) photo. We first imported the image, calibrated its size, and scaled it based on vendor-provided specifications. The 2D model was then converted to 3D by extruding to the specified thickness. The other method explored to create STL models is to utilize the digital catalog provided within Fusion 360, McMaster-Carr, which is commonly used for modeling fixation devices (e.g., screws and nails). By importing the CAD file, we finalized the STL model by finely tuning the overall dimensions. Next, we imported STL models as contours with precise shapes and sizes into RayStation (version 11A) through an in-house script. Finally, we registered them to their locations in the planning CT image set by translational and rotational adjustments. For one intracranial proton case, we created two sets of contours for the same implanted materials using the STL models and the gray-level threshold tool available in RayStation.
Results: We qualitatively assessed results by examining the fusion of the contours overlaid on the embedded hardware in CT images. Compared to the gray level-based contours, the STL-based contours significantly improved the accuracy of the reproduced geometry. This will reduce the range uncertainty associated with a proton beam going through the hardware.
Conclusion: We demonstrated that establishing an accurate STL database of the implants is feasible. This workflow can create digital models for metal- and non-metal hardware. To further utilize our STL database, we plan to 3D-print the hardware with various materials and to develop a commissioning workflow for this approach.
Keywords
Not Applicable / None Entered.
Taxonomy
Contact Email